Cognitive models of anxiety propose that maladaptive schemata guide information processing so the anxious person selectively attends to threatening cues, interprets ambiguous cues as threatening, and preferentially recalls threat cues. There is robust evidence for many of the hypothesized cognitive biases, but little is known about how these biases relate to one another, how they relate to panic symptoms, or how they change in response to treatment. This proposal seeks to examine relationships among information processing measures in panic disorder, and to examine the patterns of change in cognitive and symptom measures over the course of treatment to determine which cognitive measures will predict symptom reduction. Further, this study would provide the first evaluation of implicit associations in panic disorder and consider how these associations respond to treatment. In addition, the study would examine the importance of changing catastrophic misinterpretations of bodily sensations to effect symptom reduction. To evaluate these questions, a group of patients with panic disorder (N = 30) would receive cognitive behavioral group therapy after completing an assessment battery including cognitive, affective, physiological and behavioral measures of panic, as well as information processing measures (tasks evaluating implicit associations in memory to reflect schematic processing, and tasks to evaluate attention and interpretation biases). The assessment battery would be completed pre-treatment, at 3 time points during the 12-week treatment protocol, and at post-treatment and 12-month follow-up. Longitudinal growth curve modeling would be used :o evaluate the trajectories of change and relationships among different cognitive and symptom measures of panic, and the follow-up period would allow prediction of return of fear. Control groups include a delayed- treatment group of persons with panic disorder (N = 30) who would complete the assessment battery at matched time points to the experimental group, and a non-anxious healthy control group (N = 30). [unreadable] [unreadable]